We brief our distributed search
flowchart in Star Student Project follows: First, we
partition a query image into multiple overlapped rectangles as Star
Student Project , so that local feature extraction is completed
over multiple servers. Second, we Star Student Project
quantize the extracted local features from a query image into visual
words using the Star Student Project duplicate vocabulary in
each server respectively without inverted indexing . Third, we
tackle the issues Star Student Project of both What to
distribute and How to distribute from Star Student Project
a machine learning perspective, minimizing the overall search Star
Student Project latency as well as the computational cost Star
Student Project in traversing the inverted indexing files in
each Star Student Project server.